Hey — I’ve got my proportional lining numerals all set up as zero.lf, one.lf etc. When I select [Proportional lining] in InDesign, nothing changes. Looking at the features tab it doesn’t look like the class is being automatically set up. Other opentype features like stylistic alts, fractions etc are working.
The numerals are incidentally being recognised as lining figures in the left hand categories area.
You also need to have standard numerals without suffixes, otherwise there’s nothing the feature can substitute.
And do the standard figures all have the same width and height?
The default figures? No they don’t. I would think it doesn’t matter what they look like in terms of switching class?
How they are spaced does matter. Can you confirm your default numerals are of the same width?
the feature code calculation needs to know what your default figures are and tries to determine it by looking at certain dimensions. If they don’t fit or are unconventional, it fails to generate the feature code. Can you post some screenshots of your default and of you lining figure?
The default numerals are not monospaced. Basically my default figures are 92% cap height. I wanted to use the proportional lining class as the switch to cap height figures. I could of made these a stylistic set but bundling them in o/t figure settings seemed to make more sense. Perhaps I’m wrong on that though!
I appreciate all the help
this is what I meant with ‘unconventional’. Your approach is totally fine. But in this case, you need to write your feature code yourself.
But it is easy. Add a new feature, call it ‘lnum’.
Select all default numerals in the font view, choose ‘Copy Glyphs Names’ > ‘Space Separated’.
sub [ in the feature panel, paste the glyph names, type
] by [,
then select and copy the names of the lining figures and paste it in the feature and type
you should end up with something like this:
sub [ names of the number ] by [ names of the lining figure ];
easy like a Sonntag Morgen.
oooh one last question… would it be typical to bundle related glyphs such as currency, degree etc in here as well?
Rainer in another thread: Glyphs with .case and .lf suffixes will be added automatically to case. Perhaps you want to have slash.case, parenright.case and parenleft.case, which better fit your uppercase glyphs. You may want to copy your Germandbls to germandbls.case and so on…
How would one add the newly created lnum figures into case?
You just have to write like this:
sub dollar by dollar.case;
Or if you want to substitute with dollar.lnum, then just change to “by dollar.lnum”.
It’s very easy to manually write substitution rules like that, but if you don’t want to bother and want to automate it, you can add suffixes like dollar.lnum.case (which activates when you choose proportional numeral and all caps in InDesign).
By the way, case feature (in InDesign) only occurs when you turn on all caps from the character palette menu. Changing case from the type menu literally does that without retaining casing of the original text.
Just found this thread because I also have the problem that auto-generating
lnum does not work in my current font.
Shouldn’t Glyphs determine the default figure style from the suffixes found in the font? If there are numerals without suffix and with .lf then it can easily be determined that the plain figures are oldstyle, and what
lnum should look like. No need to measure proportions, which is a fragile approach. As I understand it, all the information the automatism needs is in the suffixes.
It works best if exactly one of the suffixes is missing. Otherwise there will be some guesswork involved. Can you send me the .glyphs file? I will have a look.
For reference, the numeral feature automation works best when:
- all figure sets you want in your font are generated,
- you have one set without a suffix (the default
one two three) and three sets with suffixes (leaving out the one for which you already have the defaults),
- each set is complete, i.e., contains all figures from 0 through 9.
So I suggest you only switch on automation after you created the glyphs, even if they are empty.
Logically, this is only possible reliably if you have a set of default figures (unsuffixed), plus three of the four suffixed sets. E.g., if your default figure set is lining proportional, you leave out
.lf because it already is covered by the default, but you do create
You are right, Rainer. It is not unambiguous only from the suffixes. With only plain and .lf numerals, Glyphs does not know whether the .lf figures are proportional lining as opposed to proportional oldstyle, or as opposed to tabular lining. The “.lf” suffix confused me, maybe it should be “.plf” instead?
How about supporting .lnum as a suffix? Then Glyphs could build the
lnum feature easily, in the same fashion it builds
ss01 based on the .ss01 suffix.
I think that is a good idea. Applying the principle to all suffixes would yield:
.posf .tosf .tlf .plf
The problem is that the lnum/onum and pnum/tnum features do one thing only: switch between either proportional/tabular or lining/oldstyle. But each figure design, of course, has to have both decided. A very valid question is what should be done in terms of non-figures, e.g., currency symbols or math operators. I could imagine a
plus.lnum that would apply to both tabular and proportional lining figures.
.lf are supposed to be proportionally, otherwise they need to be named